------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
      name:  <unnamed>
       log:  C:\Users\happy\Dropbox\Penn State\Dissertation\splits\Joo_Mukherjee_Replication\joo_mukherjee_onlineappdx.log
  log type:  text
 opened on:  13 Oct 2020, 01:17:01

. ***********************************************

. 
. * Replication Material For II Joo & Mukherjee *

. 
. *           Online Appendix Analysis          *    

. 
. * Use II_Joo_Mukherjee.dta for replication    *

. 
. ***********************************************

. 
. 
. 
. use II_Joo_Mukherjee

. 
. 
. 
. xtset rebid year
       panel variable:  rebid (unbalanced)
        time variable:  year, 1980 to 2014, but with gaps
                delta:  1 unit

. 
. 
. 
. *** Table A1: Autocratic and Democratic Rebels Analysis ****

. 
. 
. 
. xtprobit autocracy centtime cent dur_exist, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -726.37254  
Iteration 1:   log pseudolikelihood = -628.42411  
Iteration 2:   log pseudolikelihood = -627.93827  
Iteration 3:   log pseudolikelihood = -627.93815  
Iteration 4:   log pseudolikelihood = -627.93815  

Fitting full model:

rho =  0.0     log pseudolikelihood = -627.93815
rho =  0.1     log pseudolikelihood = -521.76082
rho =  0.2     log pseudolikelihood = -480.08066
rho =  0.3     log pseudolikelihood = -457.16266
rho =  0.4     log pseudolikelihood = -443.32075
rho =  0.5     log pseudolikelihood = -435.14791
rho =  0.6     log pseudolikelihood = -431.47714
rho =  0.7     log pseudolikelihood = -432.43502

Iteration 0:   log pseudolikelihood =  -428.7925  
Iteration 1:   log pseudolikelihood =  -369.0252  
Iteration 2:   log pseudolikelihood =  -355.3042  
Iteration 3:   log pseudolikelihood = -354.71459  
Iteration 4:   log pseudolikelihood = -354.69788  
Iteration 5:   log pseudolikelihood = -354.69788  (backed up)
Iteration 6:   log pseudolikelihood = -354.69408  
Iteration 7:   log pseudolikelihood = -354.69408  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,101
Group variable: rebid                           Number of groups  =        175

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =        6.3
                                                              max =         32

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(3)      =      11.51
Log pseudolikelihood  = -354.69408              Prob > chi2       =     0.0093

                                (Std. Err. adjusted for 175 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
   autocracy |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    centtime |   .1296576   .0617568     2.10   0.036     .0086164    .2506988
        cent |  -1.418208   .6804201    -2.08   0.037    -2.751807    -.084609
   dur_exist |  -.1377156   .1026581    -1.34   0.180    -.3389217    .0634905
       _cons |   -1.42445    1.04036    -1.37   0.171    -3.463519    .6146186
-------------+----------------------------------------------------------------
    /lnsig2u |   2.217819   .8034691                      .6430487     3.79259
-------------+----------------------------------------------------------------
     sigma_u |   3.031052   1.217678                      1.379229    6.661168
         rho |   .9018383   .0711279                      .6554423    .9779596
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit autocracy centtime cent dur_exist negotiation ceasefire rstrength, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -720.28967  
Iteration 1:   log pseudolikelihood = -618.01663  
Iteration 2:   log pseudolikelihood = -617.38433  
Iteration 3:   log pseudolikelihood = -617.38416  
Iteration 4:   log pseudolikelihood = -617.38416  

Fitting full model:

rho =  0.0     log pseudolikelihood = -617.38416
rho =  0.1     log pseudolikelihood = -513.97723
rho =  0.2     log pseudolikelihood = -472.65066
rho =  0.3     log pseudolikelihood = -449.64038
rho =  0.4     log pseudolikelihood = -435.33608
rho =  0.5     log pseudolikelihood = -426.51745
rho =  0.6     log pseudolikelihood =  -422.5267
rho =  0.7     log pseudolikelihood = -423.46282

Iteration 0:   log pseudolikelihood = -420.73554  
Iteration 1:   log pseudolikelihood =   -361.348  
Iteration 2:   log pseudolikelihood = -350.82277  
Iteration 3:   log pseudolikelihood = -350.40596  
Iteration 4:   log pseudolikelihood = -350.38626  
Iteration 5:   log pseudolikelihood = -350.38626  (backed up)
Iteration 6:   log pseudolikelihood = -350.37848  
Iteration 7:   log pseudolikelihood = -350.37848  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,088
Group variable: rebid                           Number of groups  =        168

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =        6.5
                                                              max =         32

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(6)      =      14.69
Log pseudolikelihood  = -350.37848              Prob > chi2       =     0.0228

                                (Std. Err. adjusted for 168 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
   autocracy |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    centtime |   .1270712   .0611854     2.08   0.038     .0071499    .2469925
        cent |   -1.40178   .6892844    -2.03   0.042    -2.752752    -.050807
   dur_exist |  -.1366585   .1025018    -1.33   0.182    -.3375584    .0642414
 negotiation |   .3327612   .2482765     1.34   0.180    -.1538517    .8193742
   ceasefire |   .5689483   .5122339     1.11   0.267    -.4350117    1.572908
   rstrength |  -.2446448   .3077132    -0.80   0.427    -.8477516     .358462
       _cons |  -1.051161   1.176517    -0.89   0.372    -3.357092     1.25477
-------------+----------------------------------------------------------------
    /lnsig2u |   2.217477   .7836026                       .681644     3.75331
-------------+----------------------------------------------------------------
     sigma_u |   3.030533   1.187367                      1.406103     6.53162
         rho |    .901808   .0693883                      .6641055    .9770968
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit autocracy centtime cent dur_exist negotiation ceasefire rstrength terr mobilize, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -708.78796  
Iteration 1:   log pseudolikelihood = -597.82389  
Iteration 2:   log pseudolikelihood = -597.12009  
Iteration 3:   log pseudolikelihood = -597.12008  

Fitting full model:

rho =  0.0     log pseudolikelihood = -597.12008
rho =  0.1     log pseudolikelihood = -500.48619
rho =  0.2     log pseudolikelihood = -461.78458
rho =  0.3     log pseudolikelihood = -440.20652
rho =  0.4     log pseudolikelihood = -427.01816
rho =  0.5     log pseudolikelihood = -419.46649
rho =  0.6     log pseudolikelihood = -416.07217
rho =  0.7     log pseudolikelihood = -416.20137

Iteration 0:   log pseudolikelihood = -413.44888  
Iteration 1:   log pseudolikelihood = -354.85252  
Iteration 2:   log pseudolikelihood = -344.69092  
Iteration 3:   log pseudolikelihood = -344.23841  
Iteration 4:   log pseudolikelihood = -344.21625  
Iteration 5:   log pseudolikelihood = -344.21625  (backed up)
Iteration 6:   log pseudolikelihood =  -344.2082  
Iteration 7:   log pseudolikelihood =  -344.2082  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,069
Group variable: rebid                           Number of groups  =        159

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =        6.7
                                                              max =         32

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(8)      =      15.21
Log pseudolikelihood  =  -344.2082              Prob > chi2       =     0.0552

                                (Std. Err. adjusted for 159 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
   autocracy |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    centtime |   .1288858   .0621172     2.07   0.038     .0071383    .2506334
        cent |  -1.496977   .7573337    -1.98   0.048    -2.981324   -.0126305
   dur_exist |  -.1409802   .1036115    -1.36   0.174    -.3440551    .0620947
 negotiation |   .3463036   .2521707     1.37   0.170    -.1479418     .840549
   ceasefire |   .5734267   .5189437     1.10   0.269    -.4436842    1.590538
   rstrength |  -.3747623   .3897405    -0.96   0.336     -1.13864     .389115
        terr |  -.1461165   .6820025    -0.21   0.830    -1.482817    1.190584
    mobilize |   .7798077   .5229154     1.49   0.136    -.2450877    1.804703
       _cons |   -1.67562   1.464751    -1.14   0.253     -4.54648    1.195239
-------------+----------------------------------------------------------------
    /lnsig2u |    2.21699   .7866011                        .67528      3.7587
-------------+----------------------------------------------------------------
     sigma_u |   3.029795    1.19162                      1.401636    6.549246
         rho |   .9017649   .0696811                      .6626844    .9772171
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit autocracy centtime cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -541.64537  
Iteration 1:   log pseudolikelihood = -432.30143  
Iteration 2:   log pseudolikelihood = -431.07589  
Iteration 3:   log pseudolikelihood = -431.07363  
Iteration 4:   log pseudolikelihood = -431.07363  

Fitting full model:

rho =  0.0     log pseudolikelihood = -431.07363
rho =  0.1     log pseudolikelihood = -372.82253
rho =  0.2     log pseudolikelihood = -344.23292
rho =  0.3     log pseudolikelihood = -326.16162
rho =  0.4     log pseudolikelihood = -313.57425
rho =  0.5     log pseudolikelihood = -304.71556
rho =  0.6     log pseudolikelihood = -298.80089
rho =  0.7     log pseudolikelihood = -295.37505
rho =  0.8     log pseudolikelihood = -297.04339

Iteration 0:   log pseudolikelihood = -294.05208  
Iteration 1:   log pseudolikelihood = -258.15776  
Iteration 2:   log pseudolikelihood = -243.49915  
Iteration 3:   log pseudolikelihood = -242.93637  (not concave)
Iteration 4:   log pseudolikelihood = -241.87559  (not concave)
Iteration 5:   log pseudolikelihood = -241.57119  (not concave)
Iteration 6:   log pseudolikelihood =  -241.4304  (not concave)
Iteration 7:   log pseudolikelihood = -241.14026  
Iteration 8:   log pseudolikelihood = -241.12773  (not concave)
Iteration 9:   log pseudolikelihood = -241.04053  (not concave)
Iteration 10:  log pseudolikelihood = -240.96961  (not concave)
Iteration 11:  log pseudolikelihood = -240.94631  (not concave)
Iteration 12:  log pseudolikelihood = -240.84948  (not concave)
Iteration 13:  log pseudolikelihood = -240.82891  (not concave)
Iteration 14:  log pseudolikelihood = -240.82891  (not concave)
Iteration 15:  log pseudolikelihood = -240.75552  (not concave)
Iteration 16:  log pseudolikelihood = -240.69855  (not concave)
Iteration 17:  log pseudolikelihood = -240.66921  (not concave)
Iteration 18:  log pseudolikelihood = -240.63844  (not concave)
Iteration 19:  log pseudolikelihood = -240.62615  (not concave)
Iteration 20:  log pseudolikelihood = -240.61739  (not concave)
Iteration 21:  log pseudolikelihood = -240.60984  (not concave)
Iteration 22:  log pseudolikelihood = -240.60315  (not concave)
Iteration 23:  log pseudolikelihood =   -240.599  (not concave)
Iteration 24:  log pseudolikelihood = -240.59613  (not concave)
Iteration 25:  log pseudolikelihood =  -240.5938  (not concave)
Iteration 26:  log pseudolikelihood = -240.59194  (not concave)
Iteration 27:  log pseudolikelihood = -240.59043  (not concave)
Iteration 28:  log pseudolikelihood = -240.58922  (not concave)
Iteration 29:  log pseudolikelihood = -240.58824  (not concave)
Iteration 30:  log pseudolikelihood = -240.58742  
Iteration 31:  log pseudolikelihood = -240.58382  
Iteration 32:  log pseudolikelihood = -240.58264  
Iteration 33:  log pseudolikelihood = -240.58264  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =        801
Group variable: rebid                           Number of groups  =        156

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =        5.1
                                                              max =         23

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(11)     =      26.87
Log pseudolikelihood  = -240.58264              Prob > chi2       =     0.0048

                                (Std. Err. adjusted for 156 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
   autocracy |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    centtime |   .1606929   .0446025     3.60   0.000     .0732736    .2481122
        cent |    -1.4581   .9776818    -1.49   0.136    -3.374321     .458121
   dur_exist |   -.212611   .0736764    -2.89   0.004    -.3570141   -.0682079
 negotiation |   .0333782   .3270573     0.10   0.919    -.6076423    .6743987
   ceasefire |   1.150156   .6305826     1.82   0.068    -.0857632    2.386075
   rstrength |  -1.360216   .9694202    -1.40   0.161    -3.260245    .5398125
        terr |   .5077653   1.742298     0.29   0.771    -2.907075    3.922606
    mobilize |   1.503056   1.211528     1.24   0.215    -.8714954    3.877607
independence |   3.232901   .9682796     3.34   0.001     1.335108    5.130694
       logbd |   .0886814   .0671159     1.32   0.186    -.0428634    .2202262
      logbd2 |   .0060291    .007042     0.86   0.392    -.0077729    .0198311
       _cons |  -4.514836   2.747721    -1.64   0.100     -9.90027    .8705981
-------------+----------------------------------------------------------------
    /lnsig2u |   3.499484   .2224375                      3.063514    3.935453
-------------+----------------------------------------------------------------
     sigma_u |   5.753118   .6398546                      4.626299    7.154394
         rho |   .9706731   .0063321                      .9553624    .9808375
------------------------------------------------------------------------------

. 
. 
. 
. 
. 
. *** Table A4:  binary centralization strength variable ***

. 
. xtprobit split centtime_bi cent dur_exist t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -609.61334  
Iteration 1:   log pseudolikelihood = -596.80036  
Iteration 2:   log pseudolikelihood = -596.27488  
Iteration 3:   log pseudolikelihood = -596.27283  
Iteration 4:   log pseudolikelihood = -596.27283  

Fitting full model:

rho =  0.0     log pseudolikelihood = -596.27283
rho =  0.1     log pseudolikelihood = -591.98462
rho =  0.2     log pseudolikelihood = -596.04532

Iteration 0:   log pseudolikelihood = -591.98469  
Iteration 1:   log pseudolikelihood = -590.54983  
Iteration 2:   log pseudolikelihood = -590.52477  
Iteration 3:   log pseudolikelihood = -590.52462  
Iteration 4:   log pseudolikelihood = -590.52462  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,725
Group variable: rebid                           Number of groups  =        240

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       15.5
                                                              max =         35

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(6)      =      15.14
Log pseudolikelihood  = -590.52462              Prob > chi2       =     0.0192

                                (Std. Err. adjusted for 240 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
 centtime_bi |   .0247035   .0065041     3.80   0.000     .0119557    .0374512
        cent |  -.2018449   .0831365    -2.43   0.015    -.3647895   -.0389003
   dur_exist |  -.0157253   .0071998    -2.18   0.029    -.0298367    -.001614
          t1 |   30.33621   39.25918     0.77   0.440    -46.61037    107.2828
          t2 |  -.5625863   2.739958    -0.21   0.837    -5.932806    4.807633
          t3 |  -.0088405   .0572674    -0.15   0.877    -.1210825    .1034016
       _cons |  -1.492181   .2060088    -7.24   0.000    -1.895951   -1.088412
-------------+----------------------------------------------------------------
    /lnsig2u |  -2.037138   .4129939                     -2.846591   -1.227684
-------------+----------------------------------------------------------------
     sigma_u |   .3611114   .0745684                      .2409188    .5412672
         rho |   .1153585   .0421464                      .0548578     .226587
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime_bi cent dur_exist negotiation ceasefire rstrength t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -576.60311  
Iteration 1:   log pseudolikelihood = -539.52079  
Iteration 2:   log pseudolikelihood = -537.51908  
Iteration 3:   log pseudolikelihood = -537.50962  
Iteration 4:   log pseudolikelihood = -537.50962  

Fitting full model:

rho =  0.0     log pseudolikelihood = -537.50962
rho =  0.1     log pseudolikelihood = -539.51559

Iteration 0:   log pseudolikelihood = -539.51556  
Iteration 1:   log pseudolikelihood =  -536.9827  
Iteration 2:   log pseudolikelihood = -536.81552  
Iteration 3:   log pseudolikelihood = -536.80863  
Iteration 4:   log pseudolikelihood = -536.80861  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,382
Group variable: rebid                           Number of groups  =        220

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       15.4
                                                              max =         34

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      65.46
Log pseudolikelihood  = -536.80861              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 220 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .3770684   .1601995     2.35   0.019     .0630831    .6910537
 centtime_bi |   .0225245   .0056513     3.99   0.000     .0114482    .0336008
        cent |  -.2490643    .077931    -3.20   0.001    -.4018063   -.0963223
   dur_exist |  -.0125085   .0061077    -2.05   0.041    -.0244793   -.0005376
 negotiation |   .2514862   .0995254     2.53   0.012     .0564199    .4465524
   ceasefire |   .3903374   .2562147     1.52   0.128    -.1118342    .8925091
   rstrength |   .3084856   .0657769     4.69   0.000     .1795653    .4374059
          t1 |  -37.08075   46.57223    -0.80   0.426    -128.3606    54.19913
          t2 |   3.614137   3.217492     1.12   0.261    -2.692031    9.920305
          t3 |  -.0850675   .0648418    -1.31   0.190    -.2121551    .0420201
       _cons |  -1.667068   .2457346    -6.78   0.000    -2.148699   -1.185437
-------------+----------------------------------------------------------------
    /lnsig2u |   -3.15543    .863142                     -4.847157   -1.463703
-------------+----------------------------------------------------------------
     sigma_u |   .2064463   .0890962                       .088604    .4810176
         rho |   .0408778   .0338411                      .0077895    .1879016
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime_bi cent dur_exist negotiation ceasefire rstrength terr mobilize t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -564.78739  
Iteration 1:   log pseudolikelihood = -526.38119  
Iteration 2:   log pseudolikelihood =  -524.3217  
Iteration 3:   log pseudolikelihood = -524.31212  
Iteration 4:   log pseudolikelihood = -524.31212  

Fitting full model:

rho =  0.0     log pseudolikelihood = -524.31212
rho =  0.1     log pseudolikelihood = -526.75932

Iteration 0:   log pseudolikelihood = -526.75922  
Iteration 1:   log pseudolikelihood = -524.11839  
Iteration 2:   log pseudolikelihood = -523.88792  
Iteration 3:   log pseudolikelihood = -523.86944  
Iteration 4:   log pseudolikelihood = -523.86924  
Iteration 5:   log pseudolikelihood = -523.86924  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,251
Group variable: rebid                           Number of groups  =        209

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       15.6
                                                              max =         34

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(12)     =      72.36
Log pseudolikelihood  = -523.86924              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 209 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .3833068   .1592571     2.41   0.016     .0711686    .6954449
 centtime_bi |   .0236497   .0056872     4.16   0.000     .0125031    .0347964
        cent |  -.2724559   .0813493    -3.35   0.001    -.4318977   -.1130141
   dur_exist |  -.0110986   .0062129    -1.79   0.074    -.0232756    .0010784
 negotiation |   .2721107   .1005364     2.71   0.007      .075063    .4691584
   ceasefire |   .4236297   .2540859     1.67   0.095    -.0743696    .9216289
   rstrength |   .3365784   .0697124     4.83   0.000     .1999446    .4732123
        terr |  -.1511113   .1088727    -1.39   0.165    -.3644979    .0622752
    mobilize |   -.032545   .0910992    -0.36   0.721    -.2110962    .1460061
          t1 |  -43.12602   47.56441    -0.91   0.365    -136.3505     50.0985
          t2 |     3.8061   3.278116     1.16   0.246    -2.618889    10.23109
          t3 |  -.0868676   .0659139    -1.32   0.188    -.2160565    .0423213
       _cons |  -1.546871   .2560647    -6.04   0.000    -2.048748   -1.044993
-------------+----------------------------------------------------------------
    /lnsig2u |  -3.388868   1.071357                      -5.48869   -1.289047
-------------+----------------------------------------------------------------
     sigma_u |   .1837031   .0984058                      .0642904    .5249126
         rho |   .0326452   .0338329                      .0041162    .2160142
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime_bi cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2 t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -477.43893  
Iteration 1:   log pseudolikelihood = -436.38385  
Iteration 2:   log pseudolikelihood = -433.94247  
Iteration 3:   log pseudolikelihood = -433.92994  
Iteration 4:   log pseudolikelihood = -433.92994  

Fitting full model:

rho =  0.0     log pseudolikelihood = -433.92994
rho =  0.1     log pseudolikelihood = -436.92287

Iteration 0:   log pseudolikelihood = -436.92284  
Iteration 1:   log pseudolikelihood = -434.37263  
Iteration 2:   log pseudolikelihood = -433.99299  
Iteration 3:   log pseudolikelihood = -433.90433  
Iteration 4:   log pseudolikelihood = -433.89393  
Iteration 5:   log pseudolikelihood = -433.89388  
Iteration 6:   log pseudolikelihood = -433.89388  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      2,479
Group variable: rebid                           Number of groups  =        191

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       13.0
                                                              max =         26

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(15)     =     106.78
Log pseudolikelihood  = -433.89388              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 191 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .4599323   .1535134     3.00   0.003     .1590516    .7608131
 centtime_bi |   .0226506    .005476     4.14   0.000      .011918    .0333833
        cent |  -.2315072    .091021    -2.54   0.011    -.4099051   -.0531093
   dur_exist |  -.0067388   .0070865    -0.95   0.342    -.0206282    .0071506
 negotiation |   .2543832   .1072514     2.37   0.018     .0441743    .4645921
   ceasefire |      .2854   .2590159     1.10   0.271    -.2222619    .7930618
   rstrength |   .2733341   .0830165     3.29   0.001     .1106247    .4360434
        terr |   -.192702   .1154376    -1.67   0.095    -.4189556    .0335515
    mobilize |   .0152372   .0956506     0.16   0.873    -.1722345    .2027089
independence |   -.243584    .118603    -2.05   0.040    -.4760415   -.0111264
       logbd |  -.0258467   .0195626    -1.32   0.186    -.0641888    .0124953
      logbd2 |  -.0040973   .0022785    -1.80   0.072    -.0085631    .0003685
          t1 |  -33.74615   51.14825    -0.66   0.509    -133.9949    66.50258
          t2 |   2.229923   3.361859     0.66   0.507    -4.359199    8.819045
          t3 |  -.0520865   .0653592    -0.80   0.425    -.1801883    .0760152
       _cons |  -1.087432   .3227241    -3.37   0.001    -1.719959   -.4549043
-------------+----------------------------------------------------------------
    /lnsig2u |  -4.515286   3.272166                     -10.92861    1.898042
-------------+----------------------------------------------------------------
     sigma_u |   .1045967    .171129                      .0042353     2.58318
         rho |   .0108221   .0350284                      .0000179    .8696698
------------------------------------------------------------------------------

. 
. 
. 
. 
. 
. *** Table A5: civil-conflict years sample ***

. 
. xtprobit split centtime cent dur_exist t1 t2 t3 if conflict==1, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -279.89372  
Iteration 1:   log pseudolikelihood = -271.02336  
Iteration 2:   log pseudolikelihood = -270.47058  
Iteration 3:   log pseudolikelihood = -270.47022  
Iteration 4:   log pseudolikelihood = -270.47022  

Fitting full model:

rho =  0.0     log pseudolikelihood = -270.47022
rho =  0.1     log pseudolikelihood =   -270.476

Iteration 0:   log pseudolikelihood = -270.47605  
Iteration 1:   log pseudolikelihood = -269.65072  
Iteration 2:   log pseudolikelihood = -269.64841  
Iteration 3:   log pseudolikelihood = -269.64841  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,439
Group variable: rebid                           Number of groups  =        231

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =        6.2
                                                              max =         35

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(6)      =      15.22
Log pseudolikelihood  = -269.64841              Prob > chi2       =     0.0186

                                (Std. Err. adjusted for 231 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    centtime |    .024026   .0078771     3.05   0.002     .0085872    .0394647
        cent |  -.4220749   .1565629    -2.70   0.007    -.7289325   -.1152173
   dur_exist |  -.0609174   .0214092    -2.85   0.004    -.1028786   -.0189562
          t1 |   98.51662   61.12186     1.61   0.107    -21.28003    218.3133
          t2 |  -4.472046   5.030377    -0.89   0.374     -14.3314    5.387311
          t3 |   .0499944    .119206     0.42   0.675     -.183645    .2836338
       _cons |  -1.149841   .3553236    -3.24   0.001    -1.846263   -.4534201
-------------+----------------------------------------------------------------
    /lnsig2u |  -2.477397   1.054798                     -4.544763   -.4100306
-------------+----------------------------------------------------------------
     sigma_u |   .2897611   .1528197                      .1030664    .8146349
         rho |    .077458   .0753741                       .010511    .3989048
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength t1 t2 t3 if conflict==1, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -264.66335  
Iteration 1:   log pseudolikelihood = -240.47816  
Iteration 2:   log pseudolikelihood = -238.74703  
Iteration 3:   log pseudolikelihood = -238.73672  
Iteration 4:   log pseudolikelihood = -238.73672  

Fitting full model:

rho =  0.0     log pseudolikelihood = -238.73672
rho =  0.1     log pseudolikelihood = -241.29521

Iteration 0:   log pseudolikelihood = -241.29518  
Iteration 1:   log pseudolikelihood = -239.38648  
Iteration 2:   log pseudolikelihood = -238.93608  
Iteration 3:   log pseudolikelihood = -238.79275  
Iteration 4:   log pseudolikelihood = -238.75264  
Iteration 5:   log pseudolikelihood = -238.74066  
Iteration 6:   log pseudolikelihood = -238.73761  
Iteration 7:   log pseudolikelihood = -238.73691  
Iteration 8:   log pseudolikelihood = -238.73675  
Iteration 9:   log pseudolikelihood = -238.73674  
Iteration 10:  log pseudolikelihood = -238.73673  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,314
Group variable: rebid                           Number of groups  =        198

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =        6.6
                                                              max =         34

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      50.71
Log pseudolikelihood  = -238.73673              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 198 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .7476381   .2292727     3.26   0.001     .2982719    1.197004
    centtime |   .0260295   .0076016     3.42   0.001     .0111306    .0409284
        cent |  -.5367273    .142791    -3.76   0.000    -.8165926    -.256862
   dur_exist |  -.0636689   .0209211    -3.04   0.002    -.1046736   -.0226643
 negotiation |   .1106978   .1315671     0.84   0.400     -.147169    .3685646
   ceasefire |   .3207498   .2507518     1.28   0.201    -.1707146    .8122142
   rstrength |   .3715183     .09192     4.04   0.000     .1913583    .5516782
          t1 |    61.3472   66.71215     0.92   0.358     -69.4062    192.1006
          t2 |  -1.573012   5.303888    -0.30   0.767    -11.96844    8.822418
          t3 |  -.0086989   .1205567    -0.07   0.942    -.2449857    .2275879
       _cons |  -1.485697   .4226656    -3.52   0.000    -2.314107   -.6572881
-------------+----------------------------------------------------------------
    /lnsig2u |  -14.24981   113428.6                     -222330.3    222301.8
-------------+----------------------------------------------------------------
     sigma_u |   .0008048   45.64418                             0           .
         rho |   6.48e-07   .0734696                             0           .
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize t1 t2 t3 if conflict==1, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -263.50196  
Iteration 1:   log pseudolikelihood = -238.04337  
Iteration 2:   log pseudolikelihood = -236.28799  
Iteration 3:   log pseudolikelihood = -236.27833  
Iteration 4:   log pseudolikelihood = -236.27833  

Fitting full model:

rho =  0.0     log pseudolikelihood = -236.27833
rho =  0.1     log pseudolikelihood = -238.99684

Iteration 0:   log pseudolikelihood =  -238.9968  
Iteration 1:   log pseudolikelihood = -236.99184  
Iteration 2:   log pseudolikelihood = -236.50241  
Iteration 3:   log pseudolikelihood = -236.33917  
Iteration 4:   log pseudolikelihood = -236.29456  
Iteration 5:   log pseudolikelihood = -236.28227  
Iteration 6:   log pseudolikelihood = -236.27908  
Iteration 7:   log pseudolikelihood = -236.27847  
Iteration 8:   log pseudolikelihood = -236.27836  
Iteration 9:   log pseudolikelihood = -236.27836  
Iteration 10:  log pseudolikelihood = -236.27835  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,292
Group variable: rebid                           Number of groups  =        188

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =        6.9
                                                              max =         34

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(12)     =      56.05
Log pseudolikelihood  = -236.27835              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 188 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .7401187   .2429977     3.05   0.002     .2638521    1.216385
    centtime |   .0253278     .00736     3.44   0.001     .0109025     .039753
        cent |  -.5457508   .1446114    -3.77   0.000    -.8291839   -.2623176
   dur_exist |  -.0599915    .020302    -2.95   0.003    -.0997826   -.0202004
 negotiation |   .1491045   .1318748     1.13   0.258    -.1093653    .4075744
   ceasefire |   .3501538   .2536072     1.38   0.167    -.1469072    .8472149
   rstrength |   .4377143   .1047726     4.18   0.000     .2323637    .6430648
        terr |  -.0798209   .1343562    -0.59   0.552    -.3431543    .1835125
    mobilize |  -.1769245   .1127783    -1.57   0.117    -.3979659    .0441169
          t1 |   59.82682   67.40305     0.89   0.375    -72.28073    191.9344
          t2 |  -1.737163   5.206175    -0.33   0.739    -11.94108    8.466752
          t3 |  -.0038846   .1171415    -0.03   0.974    -.2334776    .2257085
       _cons |  -1.296724   .4340284    -2.99   0.003    -2.147404    -.446044
-------------+----------------------------------------------------------------
    /lnsig2u |  -13.51886   70850.75                     -138878.4    138851.4
-------------+----------------------------------------------------------------
     sigma_u |   .0011599   41.08954                             0           .
         rho |   1.35e-06   .0953184                             0           .
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2 t1 t2 t3 if conflict==1, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -223.83438  
Iteration 1:   log pseudolikelihood = -200.54357  
Iteration 2:   log pseudolikelihood = -198.99088  
Iteration 3:   log pseudolikelihood = -198.71141  
Iteration 4:   log pseudolikelihood = -198.12566  
Iteration 5:   log pseudolikelihood = -198.12195  
Iteration 6:   log pseudolikelihood = -198.12195  

Fitting full model:

rho =  0.0     log pseudolikelihood = -198.12195
rho =  0.1     log pseudolikelihood = -200.35246

Iteration 0:   log pseudolikelihood = -200.35244  
Iteration 1:   log pseudolikelihood = -198.70288  
Iteration 2:   log pseudolikelihood = -198.30208  
Iteration 3:   log pseudolikelihood = -198.17127  
Iteration 4:   log pseudolikelihood = -198.13182  
Iteration 5:   log pseudolikelihood = -198.12416  
Iteration 6:   log pseudolikelihood = -198.12241  
Iteration 7:   log pseudolikelihood = -198.12202  
Iteration 8:   log pseudolikelihood = -198.12196  
Iteration 9:   log pseudolikelihood = -198.12195  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =        911
Group variable: rebid                           Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =        5.8
                                                              max =         26

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(15)     =      72.70
Log pseudolikelihood  = -198.12195              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 157 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |    .785144   .2561781     3.06   0.002     .2830441    1.287244
    centtime |   .0215995   .0072451     2.98   0.003     .0073993    .0357997
        cent |  -.4511128   .1583893    -2.85   0.004    -.7615501   -.1406755
   dur_exist |  -.0391648   .0215279    -1.82   0.069    -.0813587     .003029
 negotiation |   .1130028    .142019     0.80   0.426    -.1653494    .3913549
   ceasefire |     .24926   .2627964     0.95   0.343    -.2658115    .7643315
   rstrength |   .4297826   .1321852     3.25   0.001     .1707044    .6888608
        terr |  -.1161887   .1672623    -0.69   0.487    -.4440168    .2116394
    mobilize |  -.1442897   .1363069    -1.06   0.290    -.4114463    .1228669
independence |   .0037156   .1868527     0.02   0.984    -.3625089    .3699401
       logbd |   .4386481   .3469247     1.26   0.206    -.2413118    1.118608
      logbd2 |  -.0449558    .030152    -1.49   0.136    -.1040527     .014141
          t1 |   38.92733   76.99422     0.51   0.613    -111.9786    189.8332
          t2 |  -3.246729   5.502693    -0.59   0.555    -14.03181    7.538351
          t3 |   .0403836   .1169425     0.35   0.730    -.1888194    .2695867
       _cons |  -2.193776   .9219354    -2.38   0.017    -4.000736   -.3868158
-------------+----------------------------------------------------------------
    /lnsig2u |  -14.26728   131011.7                     -256792.4    256763.9
-------------+----------------------------------------------------------------
     sigma_u |   .0007978   52.26124                             0           .
         rho |   6.37e-07    .083389                             0           .
------------------------------------------------------------------------------

. 
. 
. 
. 
. 
. *** Table A6: Additional Robustness Tests ***

. 
. * Model 21

. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2 ngroup_country polwing state_svac_reb_lag log_os_re
> b log_os_gov polity gdp_growth t1 t2 t3, re vce(cluster rebid)

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -378.45272  
Iteration 1:   log pseudolikelihood = -346.99098  
Iteration 2:   log pseudolikelihood = -345.41779  
Iteration 3:   log pseudolikelihood = -345.41118  
Iteration 4:   log pseudolikelihood = -345.41118  

Fitting full model:

rho =  0.0     log pseudolikelihood = -345.41118
rho =  0.1     log pseudolikelihood =  -348.3334

Iteration 0:   log pseudolikelihood =  -348.3334  
Iteration 1:   log pseudolikelihood = -346.04387  
Iteration 2:   log pseudolikelihood =  -345.6244  
Iteration 3:   log pseudolikelihood = -345.46957  
Iteration 4:   log pseudolikelihood = -345.42682  
Iteration 5:   log pseudolikelihood = -345.41491  
Iteration 6:   log pseudolikelihood = -345.41215  
Iteration 7:   log pseudolikelihood = -345.41141  
Iteration 8:   log pseudolikelihood = -345.41123  
Iteration 9:   log pseudolikelihood = -345.41121  
Iteration 10:  log pseudolikelihood =  -345.4112  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,925
Group variable: rebid                           Number of groups  =        169

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       11.4
                                                              max =         25

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(22)     =      94.60
Log pseudolikelihood  =  -345.4112              Prob > chi2       =     0.0000

                                      (Std. Err. adjusted for 169 clusters in rebid)
------------------------------------------------------------------------------------
                   |               Robust
             split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
         split_lag |     .46804   .1925491     2.43   0.015     .0906507    .8454292
          centtime |   .0159041   .0049343     3.22   0.001     .0062331    .0255751
              cent |  -.3012926   .1212467    -2.48   0.013    -.5389317   -.0636534
         dur_exist |   -.033812   .0116719    -2.90   0.004    -.0566885   -.0109355
       negotiation |   .2269793   .1187217     1.91   0.056    -.0057109    .4596695
         ceasefire |   .2125549   .2967598     0.72   0.474    -.3690837    .7941935
         rstrength |   .2829374   .1255188     2.25   0.024     .0369251    .5289496
              terr |   -.224318   .1337046    -1.68   0.093    -.4863742    .0377382
          mobilize |   .0619608   .1183221     0.52   0.601    -.1699461    .2938678
      independence |  -.2876265   .1395309    -2.06   0.039     -.561102   -.0141509
             logbd |  -.0260688   .0286833    -0.91   0.363     -.082287    .0301494
            logbd2 |  -.0039858   .0032071    -1.24   0.214    -.0102717    .0023001
    ngroup_country |   .0044175   .0140808     0.31   0.754    -.0231803    .0320153
           polwing |   .0250938   .1169524     0.21   0.830    -.2041286    .2543162
state_svac_reb_lag |   .1475197   .1049236     1.41   0.160    -.0581267    .3531662
        log_os_reb |   .0027799    .010098     0.28   0.783    -.0170118    .0225716
        log_os_gov |   .0041545   .0064635     0.64   0.520    -.0085137    .0168227
           polity2 |   .0041166   .0104389     0.39   0.693    -.0163433    .0245765
        gdp_growth |   .0009522   .0045625     0.21   0.835    -.0079901    .0098945
                t1 |   -24.9131   55.06246    -0.45   0.651    -132.8335    83.00734
                t2 |   1.811915   3.495798     0.52   0.604    -5.039724    8.663554
                t3 |  -.0431529   .0656329    -0.66   0.511     -.171791    .0854851
             _cons |  -1.008918     .53808    -1.88   0.061    -2.063536     .045699
-------------------+----------------------------------------------------------------
          /lnsig2u |  -12.28402   8917.007                      -17489.3    17464.73
-------------------+----------------------------------------------------------------
           sigma_u |   .0021506   9.588439                             0           .
               rho |   4.63e-06   .0412413                             0           .
------------------------------------------------------------------------------------

. 
. 
. 
. * Model 22

. 
. drop if scode==775 | scode==750
(792 observations deleted)

. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2 t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -391.05954  
Iteration 1:   log pseudolikelihood = -358.65166  
Iteration 2:   log pseudolikelihood = -356.69669  
Iteration 3:   log pseudolikelihood = -356.68287  
Iteration 4:   log pseudolikelihood = -356.68287  

Fitting full model:

rho =  0.0     log pseudolikelihood = -356.68287
rho =  0.1     log pseudolikelihood = -357.92703

Iteration 0:   log pseudolikelihood = -357.92703  
Iteration 1:   log pseudolikelihood = -356.25246  
Iteration 2:   log pseudolikelihood =  -356.1668  
Iteration 3:   log pseudolikelihood = -356.16321  
Iteration 4:   log pseudolikelihood = -356.16321  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      1,887
Group variable: rebid                           Number of groups  =        158

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       11.9
                                                              max =         26

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(15)     =      79.97
Log pseudolikelihood  = -356.16321              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 158 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .3281839   .1514514     2.17   0.030     .0313447    .6250232
    centtime |   .0166131   .0064503     2.58   0.010     .0039707    .0292554
        cent |  -.2919893   .1439696    -2.03   0.043    -.5741646    -.009814
   dur_exist |  -.0422929    .015949    -2.65   0.008    -.0735523   -.0110335
 negotiation |   .2620244   .1287673     2.03   0.042     .0096452    .5144036
   ceasefire |   .0839821    .315629     0.27   0.790    -.5346393    .7026035
   rstrength |   .2730211   .0956177     2.86   0.004     .0856139    .4604284
        terr |  -.1591127   .1426275    -1.12   0.265    -.4386573     .120432
    mobilize |  -.0031887   .1201434    -0.03   0.979    -.2386655    .2322881
independence |  -.3370225   .1503837    -2.24   0.025    -.6317692   -.0422759
       logbd |  -.0213808   .0200678    -1.07   0.287     -.060713    .0179515
      logbd2 |   -.004012   .0023776    -1.69   0.092    -.0086721     .000648
          t1 |  -35.85771   55.95587    -0.64   0.522    -145.5292    73.81378
          t2 |   3.158755   3.818886     0.83   0.408    -4.326124    10.64363
          t3 |  -.0708622   .0746633    -0.95   0.343    -.2171996    .0754751
       _cons |  -.8973071   .4466078    -2.01   0.045    -1.772642    -.021972
-------------+----------------------------------------------------------------
    /lnsig2u |  -3.022187   .9821121                     -4.947092   -1.097283
-------------+----------------------------------------------------------------
     sigma_u |   .2206685   .1083606                      .0842855    .5777341
         rho |   .0464335   .0434854                      .0070539    .2502493
------------------------------------------------------------------------------

. 
. 
. 
. clear 

. 
. 
. 
. *Model 23 

. 
. use II_Joo_Mukherjee_influence_data

. 
. 
. 
. xtset rebid year
       panel variable:  rebid (unbalanced)
        time variable:  year, 1989 to 2014, but with gaps
                delta:  1 unit

. 
. drop if devres >2
(97 observations deleted)

. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2 t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -124.96046  
Iteration 1:   log pseudolikelihood = -80.385953  
Iteration 2:   log pseudolikelihood = -62.339173  
Iteration 3:   log pseudolikelihood = -57.241747  
Iteration 4:   log pseudolikelihood = -55.645368  
Iteration 5:   log pseudolikelihood = -55.367694  
Iteration 6:   log pseudolikelihood = -55.362666  
Iteration 7:   log pseudolikelihood = -55.362664  

Fitting full model:

rho =  0.0     log pseudolikelihood = -55.362664
rho =  0.1     log pseudolikelihood = -56.009766

Iteration 0:   log pseudolikelihood = -56.009766  
Iteration 1:   log pseudolikelihood = -55.547881  
Iteration 2:   log pseudolikelihood = -55.413704  
Iteration 3:   log pseudolikelihood =  -55.37213  
Iteration 4:   log pseudolikelihood = -55.364519  
Iteration 5:   log pseudolikelihood = -55.363088  
Iteration 6:   log pseudolikelihood = -55.362765  
Iteration 7:   log pseudolikelihood = -55.362685  
Iteration 8:   log pseudolikelihood = -55.362668  
Iteration 9:   log pseudolikelihood = -55.362667  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      2,382
Group variable: rebid                           Number of groups  =        190

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       12.5
                                                              max =         26

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(15)     =      90.42
Log pseudolikelihood  = -55.362667              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 190 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   1.726717   .3933418     4.39   0.000     .9557809    2.497652
    centtime |    .102939   .0258998     3.97   0.000     .0521763    .1537017
        cent |  -2.166614   .5382509    -4.03   0.000    -3.221566   -1.111661
   dur_exist |  -.2627064   .0760415    -3.45   0.001    -.4117451   -.1136678
 negotiation |    .744862   .3132672     2.38   0.017     .1308695    1.358854
   ceasefire |   1.121724   .3921222     2.86   0.004      .353179     1.89027
   rstrength |   1.177765   .2184873     5.39   0.000     .7495382    1.605993
        terr |  -.3174199   .3247882    -0.98   0.328    -.9539931    .3191533
    mobilize |  -.2865198     .22682    -1.26   0.207    -.7310789    .1580393
independence |   -.896285   .4582759    -1.96   0.050    -1.794489    .0019192
       logbd |  -.1008753   .0525101    -1.92   0.055    -.2037932    .0020426
      logbd2 |  -.0187888   .0070725    -2.66   0.008    -.0326507   -.0049269
          t1 |  -339.8863   201.5136    -1.69   0.092    -734.8457    55.07318
          t2 |   46.32296   17.99563     2.57   0.010     11.05218    81.59374
          t3 |  -1.277156   .4410142    -2.90   0.004    -2.141528    -.412784
       _cons |   .9773958   .7669951     1.27   0.203    -.5258869    2.480678
-------------+----------------------------------------------------------------
    /lnsig2u |  -14.10905   608225.9                      -1192115     1192087
-------------+----------------------------------------------------------------
     sigma_u |   .0008635   262.5993                             0           .
         rho |   7.46e-07   .4535044                             0           .
------------------------------------------------------------------------------

. 
. 
. 
. clear 

. 
. 
. 
. *Model 24 

. 
. use II_Joo_Mukherjee_influence_data

. 
. 
. 
. xtset rebid year
       panel variable:  rebid (unbalanced)
        time variable:  year, 1989 to 2014, but with gaps
                delta:  1 unit

. 
. drop if hat >0.0123
(252 observations deleted)

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2 t1 t2 t3, re vce(cluster rebid) 
note: split_lag != 0 predicts failure perfectly
      split_lag dropped and 9 obs not used

note: ceasefire != 0 predicts failure perfectly
      ceasefire dropped and 1 obs not used


Fitting comparison model:

Iteration 0:   log pseudolikelihood = -370.16838  
Iteration 1:   log pseudolikelihood = -346.99925  
Iteration 2:   log pseudolikelihood = -346.12001  
Iteration 3:   log pseudolikelihood = -346.11708  
Iteration 4:   log pseudolikelihood = -346.11708  

Fitting full model:

rho =  0.0     log pseudolikelihood = -346.11708
rho =  0.1     log pseudolikelihood = -346.99775

Iteration 0:   log pseudolikelihood = -346.99775  
Iteration 1:   log pseudolikelihood = -345.80244  
Iteration 2:   log pseudolikelihood = -345.71698  
Iteration 3:   log pseudolikelihood = -345.71339  
Iteration 4:   log pseudolikelihood = -345.71339  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      2,217
Group variable: rebid                           Number of groups  =        178

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       12.5
                                                              max =         26

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(13)     =      54.26
Log pseudolikelihood  = -345.71339              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 178 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |          0  (omitted)
    centtime |   .0143007   .0070201     2.04   0.042     .0005416    .0280598
        cent |  -.2348072   .1365145    -1.72   0.085    -.5023707    .0327563
   dur_exist |  -.0320152   .0161041    -1.99   0.047    -.0635787   -.0004518
 negotiation |   .3490233   .1284529     2.72   0.007     .0972602    .6007865
   ceasefire |          0  (omitted)
   rstrength |   .3069155    .114067     2.69   0.007     .0833482    .5304828
        terr |  -.2234142   .1414743    -1.58   0.114    -.5006987    .0538704
    mobilize |   .0398342   .1310442     0.30   0.761    -.2170077     .296676
independence |  -.2741513     .12016    -2.28   0.023    -.5096606    -.038642
       logbd |  -.0309009   .0267753    -1.15   0.248    -.0833795    .0215777
      logbd2 |  -.0042849   .0030388    -1.41   0.159    -.0102408     .001671
          t1 |  -8.034797   57.43278    -0.14   0.889     -120.601    104.5314
          t2 |  -.1095676   3.730938    -0.03   0.977    -7.422072    7.202936
          t3 |    .007308   .0707031     0.10   0.918    -.1312676    .1458837
       _cons |  -1.243524   .4492249    -2.77   0.006    -2.123989   -.3630599
-------------+----------------------------------------------------------------
    /lnsig2u |  -3.080017   1.146628                     -5.327365    -.832668
-------------+----------------------------------------------------------------
     sigma_u |   .2143793   .1229066                      .0696911      .65946
         rho |   .0439391   .0481681                      .0048334    .3030812
------------------------------------------------------------------------------

. 
. 
. 
. clear

. 
. 
. 
. **** Table A7: Models without young groups ****

. 
. 
. 
. use II_Joo_Mukherjee

. 
. 
. 
. xtset rebid year
       panel variable:  rebid (unbalanced)
        time variable:  year, 1980 to 2014, but with gaps
                delta:  1 unit

. 
. 
. 
. drop if max_year < 18.1776 - 14.27957
(58 observations deleted)

. 
. 
. 
. xtprobit split centtime cent dur_exist t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood =  -588.0313  
Iteration 1:   log pseudolikelihood = -579.46509  
Iteration 2:   log pseudolikelihood = -579.18602  
Iteration 3:   log pseudolikelihood = -579.18592  
Iteration 4:   log pseudolikelihood = -579.18592  

Fitting full model:

rho =  0.0     log pseudolikelihood = -579.18592
rho =  0.1     log pseudolikelihood = -574.55107
rho =  0.2     log pseudolikelihood =   -578.451

Iteration 0:   log pseudolikelihood = -574.55112  
Iteration 1:   log pseudolikelihood = -573.05074  
Iteration 2:   log pseudolikelihood = -573.02341  
Iteration 3:   log pseudolikelihood = -573.02324  
Iteration 4:   log pseudolikelihood = -573.02324  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,668
Group variable: rebid                           Number of groups  =        214

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       17.1
                                                              max =         35

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(6)      =       9.88
Log pseudolikelihood  = -573.02324              Prob > chi2       =     0.1299

                                (Std. Err. adjusted for 214 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    centtime |   .0133791   .0051089     2.62   0.009     .0033659    .0233923
        cent |   -.203994   .1066011    -1.91   0.056    -.4129284    .0049403
   dur_exist |  -.0381384   .0129718    -2.94   0.003    -.0635627   -.0127141
          t1 |    39.0333   39.99693     0.98   0.329    -39.35925    117.4258
          t2 |  -.8258939   2.762423    -0.30   0.765    -6.240144    4.588356
          t3 |  -.0045364   .0569589    -0.08   0.937    -.1161738     .107101
       _cons |  -1.539005   .2568608    -5.99   0.000    -2.042443   -1.035567
-------------+----------------------------------------------------------------
    /lnsig2u |  -2.014415   .4036421                     -2.805539   -1.223291
-------------+----------------------------------------------------------------
     sigma_u |   .3652376   .0737126                       .245915    .5424576
         rho |   .1176978   .0419162                      .0570256    .2273579
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -556.56238  
Iteration 1:   log pseudolikelihood = -523.99742  
Iteration 2:   log pseudolikelihood = -522.34919  
Iteration 3:   log pseudolikelihood = -522.34236  
Iteration 4:   log pseudolikelihood = -522.34236  

Fitting full model:

rho =  0.0     log pseudolikelihood = -522.34236
rho =  0.1     log pseudolikelihood = -524.12906

Iteration 0:   log pseudolikelihood =  -524.1291  
Iteration 1:   log pseudolikelihood = -521.54873  
Iteration 2:   log pseudolikelihood = -521.41829  
Iteration 3:   log pseudolikelihood = -521.41376  
Iteration 4:   log pseudolikelihood = -521.41375  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,358
Group variable: rebid                           Number of groups  =        203

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          2
                                                              avg =       16.5
                                                              max =         34

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      61.78
Log pseudolikelihood  = -521.41375              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 203 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .3631894   .1637293     2.22   0.027     .0422858    .6840929
    centtime |   .0144741    .004686     3.09   0.002     .0052897    .0236585
        cent |  -.2855025   .0998444    -2.86   0.004    -.4811939   -.0898111
   dur_exist |  -.0370445   .0116876    -3.17   0.002    -.0599517   -.0141373
 negotiation |   .2585085   .1033561     2.50   0.012     .0559342    .4610827
   ceasefire |   .4163807    .254331     1.64   0.102    -.0820989    .9148603
   rstrength |     .32364   .0674217     4.80   0.000      .191496    .4557841
          t1 |  -6.437923   46.49492    -0.14   0.890    -97.56629    84.69045
          t2 |   1.921111   3.173583     0.61   0.545    -4.298997    8.141219
          t3 |  -.0547493   .0632141    -0.87   0.386    -.1786467    .0691481
       _cons |  -1.768296   .2836872    -6.23   0.000    -2.324312   -1.212279
-------------+----------------------------------------------------------------
    /lnsig2u |  -3.025881   .7424019                     -4.480962     -1.5708
-------------+----------------------------------------------------------------
     sigma_u |   .2202613   .0817612                      .1064073    .4559372
         rho |   .0462702   .0327617                      .0111957    .1721023
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -544.87617  
Iteration 1:   log pseudolikelihood = -511.41298  
Iteration 2:   log pseudolikelihood = -509.77256  
Iteration 3:   log pseudolikelihood = -509.76556  
Iteration 4:   log pseudolikelihood = -509.76556  

Fitting full model:

rho =  0.0     log pseudolikelihood = -509.76556
rho =  0.1     log pseudolikelihood =  -511.9612

Iteration 0:   log pseudolikelihood = -511.96118  
Iteration 1:   log pseudolikelihood =  -509.3145  
Iteration 2:   log pseudolikelihood = -509.12674  
Iteration 3:   log pseudolikelihood = -509.11884  
Iteration 4:   log pseudolikelihood =  -509.1188  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,227
Group variable: rebid                           Number of groups  =        192

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          2
                                                              avg =       16.8
                                                              max =         34

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(12)     =      69.52
Log pseudolikelihood  =  -509.1188              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 192 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .3713619   .1629594     2.28   0.023     .0519673    .6907564
    centtime |   .0154942   .0048622     3.19   0.001     .0059645     .025024
        cent |  -.3181842   .1051223    -3.03   0.002    -.5242201   -.1121483
   dur_exist |  -.0375129   .0119715    -3.13   0.002    -.0609767   -.0140491
 negotiation |   .2815676   .1052884     2.67   0.007     .0752062    .4879291
   ceasefire |   .4442286   .2523091     1.76   0.078    -.0502882    .9387454
   rstrength |   .3537335   .0692317     5.11   0.000     .2180418    .4894251
        terr |  -.1405961   .1134533    -1.24   0.215    -.3629605    .0817684
    mobilize |  -.0354871   .0958101    -0.37   0.711    -.2232714    .1522973
          t1 |  -12.61046   47.25359    -0.27   0.790    -105.2258    80.00488
          t2 |   2.146044    3.21702     0.67   0.505    -4.159199    8.451288
          t3 |  -.0574479   .0639053    -0.90   0.369       -.1827    .0678042
       _cons |  -1.633649   .2961071    -5.52   0.000    -2.214008   -1.053289
-------------+----------------------------------------------------------------
    /lnsig2u |  -3.204339    .885141                     -4.939184   -1.469495
-------------+----------------------------------------------------------------
     sigma_u |    .201459   .0891598                      .0846194    .4796267
         rho |   .0390028   .0331765                      .0071095    .1870194
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2 t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -461.44517  
Iteration 1:   log pseudolikelihood = -425.49187  
Iteration 2:   log pseudolikelihood = -423.69219  
Iteration 3:   log pseudolikelihood = -423.68491  
Iteration 4:   log pseudolikelihood = -423.68491  

Fitting full model:

rho =  0.0     log pseudolikelihood = -423.68491
rho =  0.1     log pseudolikelihood = -426.24627

Iteration 0:   log pseudolikelihood = -426.24628  
Iteration 1:   log pseudolikelihood = -423.84351  
Iteration 2:   log pseudolikelihood = -423.57297  
Iteration 3:   log pseudolikelihood = -423.53326  
Iteration 4:   log pseudolikelihood = -423.53123  
Iteration 5:   log pseudolikelihood = -423.53121  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      2,459
Group variable: rebid                           Number of groups  =        177

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       13.9
                                                              max =         26

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(15)     =      96.31
Log pseudolikelihood  = -423.53121              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 177 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .4405254   .1583969     2.78   0.005     .1300731    .7509777
    centtime |   .0143979   .0046302     3.11   0.002     .0053228    .0234731
        cent |  -.2673181   .1200799    -2.23   0.026    -.5026704   -.0319658
   dur_exist |    -.03205   .0127743    -2.51   0.012    -.0570872   -.0070128
 negotiation |   .2710264   .1129091     2.40   0.016     .0497286    .4923241
   ceasefire |   .2765408   .2566015     1.08   0.281     -.226389    .7794705
   rstrength |    .277721   .0857282     3.24   0.001     .1096968    .4457453
        terr |  -.1615675   .1206293    -1.34   0.180    -.3979966    .0748616
    mobilize |   .0154859    .104662     0.15   0.882    -.1896478    .2206195
independence |  -.2344805   .1188724    -1.97   0.049    -.4674661   -.0014949
       logbd |  -.0256152   .0203119    -1.26   0.207    -.0654259    .0141954
      logbd2 |  -.0043536    .002364    -1.84   0.066     -.008987    .0002798
          t1 |  -5.333388   50.30996    -0.11   0.916    -103.9391    93.27232
          t2 |   .8490252   3.312768     0.26   0.798    -5.643881    7.341931
          t3 |  -.0286166   .0639312    -0.45   0.654    -.1539195    .0966864
       _cons |  -1.154976   .3825061    -3.02   0.003    -1.904674   -.4052774
-------------+----------------------------------------------------------------
    /lnsig2u |   -3.76169   1.733818                      -7.15991   -.3634697
-------------+----------------------------------------------------------------
     sigma_u |   .1524612     .13217                       .027877    .8338224
         rho |   .0227164   .0384914                      .0007765    .4101199
------------------------------------------------------------------------------

. 
. 
. 
. clear

. 
. 
. 
. 
. 
. **** Table A8: Models withoutt old groups *****

. 
. 
. 
. use II_Joo_Mukherjee

. 
. 
. 
. xtset rebid year
       panel variable:  rebid (unbalanced)
        time variable:  year, 1980 to 2014, but with gaps
                delta:  1 unit

. 
. 
. 
. drop if max_year > 59
(95 observations deleted)

. 
. 
. 
. xtprobit split centtime cent dur_exist t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -589.79352  
Iteration 1:   log pseudolikelihood = -580.04464  
Iteration 2:   log pseudolikelihood = -579.83915  
Iteration 3:   log pseudolikelihood = -579.83893  
Iteration 4:   log pseudolikelihood = -579.83893  

Fitting full model:

rho =  0.0     log pseudolikelihood = -579.83893
rho =  0.1     log pseudolikelihood = -574.62572
rho =  0.2     log pseudolikelihood =   -578.019

Iteration 0:   log pseudolikelihood = -574.62573  
Iteration 1:   log pseudolikelihood = -573.05427  
Iteration 2:   log pseudolikelihood =  -573.0378  
Iteration 3:   log pseudolikelihood = -573.03775  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,630
Group variable: rebid                           Number of groups  =        236

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       15.4
                                                              max =         35

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(6)      =      12.40
Log pseudolikelihood  = -573.03775              Prob > chi2       =     0.0535

                                (Std. Err. adjusted for 236 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    centtime |   .0131823    .005857     2.25   0.024     .0017028    .0246617
        cent |  -.2284274   .1071618    -2.13   0.033    -.4384606   -.0183941
   dur_exist |  -.0445019   .0146052    -3.05   0.002    -.0731276   -.0158762
          t1 |   53.37444   37.69265     1.42   0.157     -20.5018    127.2507
          t2 |  -1.664376   2.642123    -0.63   0.529    -6.842841     3.51409
          t3 |    .016154   .0549655     0.29   0.769    -.0915763    .1238843
       _cons |  -1.474609   .2481724    -5.94   0.000    -1.961018   -.9882006
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.925329   .3819379                     -2.673914   -1.176745
-------------+----------------------------------------------------------------
     sigma_u |   .3818739   .0729261                      .2626437    .5552302
         rho |   .1272684   .0424223                      .0645303     .235638
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -560.18249  
Iteration 1:   log pseudolikelihood = -525.58303  
Iteration 2:   log pseudolikelihood = -524.04755  
Iteration 3:   log pseudolikelihood = -524.03997  
Iteration 4:   log pseudolikelihood = -524.03997  

Fitting full model:

rho =  0.0     log pseudolikelihood = -524.03997
rho =  0.1     log pseudolikelihood = -525.04064

Iteration 0:   log pseudolikelihood = -525.04071  
Iteration 1:   log pseudolikelihood =  -522.7693  
Iteration 2:   log pseudolikelihood = -522.71242  
Iteration 3:   log pseudolikelihood = -522.71173  
Iteration 4:   log pseudolikelihood = -522.71173  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,291
Group variable: rebid                           Number of groups  =        216

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       15.2
                                                              max =         34

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      61.30
Log pseudolikelihood  = -522.71173              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 216 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .3257654    .166542     1.96   0.050    -.0006508    .6521817
    centtime |   .0135868    .005343     2.54   0.011     .0031147    .0240589
        cent |   -.300666   .1011605    -2.97   0.003    -.4989369   -.1023952
   dur_exist |  -.0403361   .0138571    -2.91   0.004    -.0674955   -.0131766
 negotiation |   .2525575   .1065011     2.37   0.018     .0438192    .4612958
   ceasefire |   .3927999   .2575804     1.52   0.127    -.1120485    .8976483
   rstrength |   .3332358   .0685135     4.86   0.000     .1989518    .4675199
          t1 |  -22.25032   46.89908    -0.47   0.635    -114.1708    69.67019
          t2 |   2.891444   3.198019     0.90   0.366    -3.376558    9.159446
          t3 |   -.067368   .0641019    -1.05   0.293    -.1930055    .0582695
       _cons |  -1.625378   .2886435    -5.63   0.000    -2.191109   -1.059647
-------------+----------------------------------------------------------------
    /lnsig2u |   -2.79949   .6287391                     -4.031796   -1.567184
-------------+----------------------------------------------------------------
     sigma_u |   .2466598   .0775423                      .1332007    .4567623
         rho |   .0573517   .0339912                      .0174331    .1726182
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -548.37318  
Iteration 1:   log pseudolikelihood = -512.52458  
Iteration 2:   log pseudolikelihood = -510.94907  
Iteration 3:   log pseudolikelihood = -510.94158  
Iteration 4:   log pseudolikelihood = -510.94158  

Fitting full model:

rho =  0.0     log pseudolikelihood = -510.94158
rho =  0.1     log pseudolikelihood = -512.39674

Iteration 0:   log pseudolikelihood = -512.39674  
Iteration 1:   log pseudolikelihood = -510.09141  
Iteration 2:   log pseudolikelihood = -509.99111  
Iteration 3:   log pseudolikelihood = -509.98843  
Iteration 4:   log pseudolikelihood = -509.98843  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      3,160
Group variable: rebid                           Number of groups  =        205

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       15.4
                                                              max =         34

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(12)     =      66.78
Log pseudolikelihood  = -509.98843              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 205 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .3312823   .1657388     2.00   0.046     .0064402    .6561244
    centtime |   .0142494   .0054281     2.63   0.009     .0036105    .0248884
        cent |  -.3263469   .1076295    -3.03   0.002    -.5372968   -.1153971
   dur_exist |  -.0398935   .0143096    -2.79   0.005    -.0679398   -.0118471
 negotiation |   .2731591   .1082036     2.52   0.012      .061084    .4852341
   ceasefire |    .424478   .2545881     1.67   0.095    -.0745055    .9234614
   rstrength |   .3617333   .0723318     5.00   0.000     .2199655    .5035011
        terr |  -.1627766   .1161227    -1.40   0.161    -.3903728    .0648197
    mobilize |  -.0197099   .0969637    -0.20   0.839    -.2097553    .1703355
          t1 |    -28.889   47.75229    -0.60   0.545    -122.4818    64.70376
          t2 |   3.122874    3.24863     0.96   0.336    -3.244324    9.490072
          t3 |  -.0699987   .0649234    -1.08   0.281    -.1972463    .0572489
       _cons |  -1.514594   .3030007    -5.00   0.000    -2.108464   -.9207231
-------------+----------------------------------------------------------------
    /lnsig2u |  -2.963889   .7378085                     -4.409967    -1.51781
-------------+----------------------------------------------------------------
     sigma_u |   .2271955   .0838134                      .1102524    .4681787
         rho |   .0490842   .0344372                      .0120096    .1797842
------------------------------------------------------------------------------

. 
. 
. 
. xtprobit split split_lag centtime cent dur_exist negotiation ceasefire rstrength terr mobilize independence logbd logbd2 t1 t2 t3, re vce(cluster rebid) 

Fitting comparison model:

Iteration 0:   log pseudolikelihood = -462.58115  
Iteration 1:   log pseudolikelihood = -424.67643  
Iteration 2:   log pseudolikelihood = -422.85039  
Iteration 3:   log pseudolikelihood = -422.84286  
Iteration 4:   log pseudolikelihood = -422.84286  

Fitting full model:

rho =  0.0     log pseudolikelihood = -422.84286
rho =  0.1     log pseudolikelihood = -424.71306

Iteration 0:   log pseudolikelihood = -424.71306  
Iteration 1:   log pseudolikelihood = -422.64108  
Iteration 2:   log pseudolikelihood = -422.48482  
Iteration 3:   log pseudolikelihood = -422.47438  
Iteration 4:   log pseudolikelihood = -422.47431  
Iteration 5:   log pseudolikelihood = -422.47431  

Calculating robust standard errors:

Random-effects probit regression                Number of obs     =      2,420
Group variable: rebid                           Number of groups  =        187

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       12.9
                                                              max =         26

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(15)     =      93.48
Log pseudolikelihood  = -422.47431              Prob > chi2       =     0.0000

                                (Std. Err. adjusted for 187 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
       split |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   split_lag |   .3979079   .1625419     2.45   0.014     .0793316    .7164841
    centtime |   .0149362   .0059314     2.52   0.012      .003311    .0265614
        cent |  -.3040743   .1290775    -2.36   0.018    -.5570615   -.0510871
   dur_exist |  -.0356706   .0159366    -2.24   0.025    -.0669058   -.0044354
 negotiation |   .2684994   .1164105     2.31   0.021      .040339    .4966599
   ceasefire |   .2762939   .2578237     1.07   0.284    -.2290313    .7816191
   rstrength |   .3016294   .0884059     3.41   0.001      .128357    .4749018
        terr |  -.2007281   .1255234    -1.60   0.110    -.4467495    .0452932
    mobilize |   .0081949   .1062814     0.08   0.939    -.2001128    .2165026
independence |  -.2236763    .121366    -1.84   0.065    -.4615493    .0141967
       logbd |  -.0191073   .0196712    -0.97   0.331    -.0576621    .0194474
      logbd2 |  -.0034347   .0023098    -1.49   0.137    -.0079619    .0010925
          t1 |  -17.92611   50.67101    -0.35   0.724    -117.2395    81.38724
          t2 |   1.502954   3.329791     0.45   0.652    -5.023317    8.029225
          t3 |  -.0356924   .0648532    -0.55   0.582    -.1628023    .0914176
       _cons |  -1.086219   .4049505    -2.68   0.007    -1.879908    -.292531
-------------+----------------------------------------------------------------
    /lnsig2u |  -3.270077   1.121997                      -5.46915   -1.071003
-------------+----------------------------------------------------------------
     sigma_u |   .1949449   .1093638                      .0649216    .5853757
         rho |   .0366121   .0395747                      .0041971    .2552124
------------------------------------------------------------------------------

. 
. 
. 
. 
. 
. clear

. 
. 
. 
. 
. 
. **** Table A11: Relationship between rebel age and command centralization ****

. 
. 
. 
. use II_Joo_Mukherjee

. 
. 
. 
. xtset rebid year
       panel variable:  rebid (unbalanced)
        time variable:  year, 1980 to 2014, but with gaps
                delta:  1 unit

. 
. 
. 
. oprobit cent dur_exist, cluster(rebid)

Iteration 0:   log pseudolikelihood = -3807.1009  
Iteration 1:   log pseudolikelihood = -3806.9586  
Iteration 2:   log pseudolikelihood = -3806.9586  

Ordered probit regression                       Number of obs     =      3,725
                                                Wald chi2(1)      =       0.02
                                                Prob > chi2       =     0.8869
Log pseudolikelihood = -3806.9586               Pseudo R2         =     0.0000

                                (Std. Err. adjusted for 240 clusters in rebid)
------------------------------------------------------------------------------
             |               Robust
        cent |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   dur_exist |  -.0006814   .0047916    -0.14   0.887    -.0100728    .0087101
-------------+----------------------------------------------------------------
       /cut1 |  -.6089791   .1242843                     -.8525718   -.3653864
       /cut2 |   .8045222   .1285898                      .5524909    1.056554
------------------------------------------------------------------------------

. 
. xtreg dur_exist cent, mle

Fitting constant-only model:
Iteration 0:   log likelihood = -13258.401
Iteration 1:   log likelihood = -13244.112
Iteration 2:   log likelihood = -13235.823
Iteration 3:   log likelihood = -13235.172
Iteration 4:   log likelihood = -13235.165

Fitting full model:
Iteration 0:   log likelihood = -13258.296
Iteration 1:   log likelihood = -13243.872
Iteration 2:   log likelihood = -13235.582
Iteration 3:   log likelihood = -13234.929
Iteration 4:   log likelihood = -13234.922

Random-effects ML regression                    Number of obs     =      3,725
Group variable: rebid                           Number of groups  =        240

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          1
                                                              avg =       15.5
                                                              max =         35

                                                LR chi2(1)        =       0.49
Log likelihood  = -13234.922                    Prob > chi2       =     0.4856

------------------------------------------------------------------------------
   dur_exist |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
        cent |  -.5290779   .7590351    -0.70   0.486    -2.016759    .9586036
       _cons |    13.8428   1.642524     8.43   0.000     10.62351    17.06209
-------------+----------------------------------------------------------------
    /sigma_u |    11.7096   .5579581                      10.66553    12.85587
    /sigma_e |   7.587456   .0908148                      7.411534    7.767554
         rho |    .704293   .0205038                      .6629198    .7431642
------------------------------------------------------------------------------
LR test of sigma_u=0: chibar2(01) = 3915.60            Prob >= chibar2 = 0.000

. 
. 
. 
. 
. 
. *** End of do File ***

. 
. 
. log close
      name:  <unnamed>
       log:  C:\Users\happy\Dropbox\Penn State\Dissertation\splits\Joo_Mukherjee_Replication\joo_mukherjee_onlineappdx.log
  log type:  text
 closed on:  13 Oct 2020, 01:20:46
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
